Update code and documentation for new name. Add README.

.github/workflows/tag.yml

@@ -102,7 +102,7 @@ jobs:
         shell: bash
         run: |
           # Replace with the name of your binary
-          binary_name="oxi_tarql"
+          binary_name="oxi_gen"
 
           dirname="$binary_name-${{ env.RELEASE_VERSION }}-${{ matrix.target }}"
           mkdir "$dirname"

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
-name = "oxi_tarql"
-version = "0.3.1"
+name = "oxi_gen"
+version = "0.4.0"
 edition = "2024"
 
 [dependencies]

README.md

@@ -0,0 +1,58 @@
+# oxi-gen
+
+A high-performance command-line tool for converting CSV/TSV files to RDF (Turtle or N-Triples) using SPARQL CONSTRUCT queries. Inspired by [Tarql](https://tarql.github.io/), oxi-gen is built in Rust on top of the [Oxigraph](https://github.com/oxigraph/oxigraph) stack and leverages multi-threaded processing to handle large datasets efficiently.
+
+Each row of the input CSV is bound as SPARQL variable substitutions and evaluated against a CONSTRUCT query, producing RDF output. Column headers become variable names (e.g., a `name` column is available as `?name`), and the special variable `?ROWNUM` holds the current row index.
+
+## Command-Line Options
+
+```
+oxi_gen -q <QUERY> [OPTIONS]
+```
+
+| Option | Short | Description |
+|---|---|---|
+| `--query <FILE>` | `-q` | SPARQL CONSTRUCT query file to apply (required) |
+| `--input <FILE>` | `-i` | Input CSV file. Omit to read from STDIN |
+| `--output <FILE>` | `-o` | Output file. Omit to write to STDOUT |
+| `--delimiter <CHAR>` | `-d` | CSV delimiter character (default: `,`) |
+| `--tab` | `-t` | Treat input as tab-separated (TSV) |
+| `--no-header-row` | `-H` | Input has no header row; columns are named `a`–`z`, `A`–`Z` |
+| `--normalize` | `-n` | Normalize column names to UPPERCASE |
+| `--escape_char <CHAR>` | `-p` | Escape character (default: `\`) |
+| `--quote_char <CHAR>` | | Quote character (default: `"`) |
+| `--ntriples` | | Output N-Triples instead of Turtle |
+| `--gzip` | `-g` | Gzip the output (requires `--output`) |
+| `--dedup[=N]` | | Deduplicate triples within a sliding window (default window: 1000, range: 1000–5000000) |
+| `--test[=N]` | | Process only the first N rows for testing (default: 5, max: 49) |
+| `--split <ORIGINAL> <SPLIT> <DELIMITER>` | | Split column ORIGINAL on DELIMITER, binding each value to SPLIT. Can be repeated |
+| `--bind-empty-strings` | | Bind empty CSV values as empty string literals instead of skipping them |
+
+## Custom SPARQL Functions
+
+oxi-gen registers two custom functions under the `tarql:` prefix (`https://semanticarts.com/tarql/`):
+
+- **`tarql:expandPrefix(?prefix)`** — returns the IRI for a given prefix name declared in the query.
+- **`tarql:expandPrefixedName(?qname)`** — expands a prefixed name (e.g., `"foaf:name"`) into a full IRI node.
+
+## Building from Source
+
+### Prerequisites
+
+- [Rust](https://www.rust-lang.org/tools/install) (1.85+ required for edition 2024)
+
+### Build a release binary
+
+```sh
+git clone git@github.com:semanticarts/oxi-gen.git
+cd oxi-gen
+cargo build --release
+```
+
+The optimized binary will be at `target/release/oxi_gen`. The release profile is configured with LTO, single codegen unit, and abort-on-panic for maximum performance.
+
+### Run directly with Cargo
+
+```sh
+cargo run --release -- -q query.sparql -i data.csv -o output.ttl
+```

src/lib.rs

@@ -23,7 +23,7 @@ use clap::{Arg, ArgAction, ArgMatches, command, value_parser};
 
 #[allow(dead_code)]
 #[derive(Default)]
-pub struct OxiTarql {
+pub struct OxiGen {
     pub delimiter: String,
     pub tab: bool,
     pub test: u32,
@@ -41,7 +41,7 @@ pub struct OxiTarql {
     pub split: Vec<(String, String, String)>,
 }
 
-impl OxiTarql {
+impl OxiGen {
     pub fn transform(&mut self) -> Result<(), Box<dyn Error>> {
         let num_workers: usize = num_cpus::get();
 
@@ -424,7 +424,7 @@ where
     I: IntoIterator<Item = String>,
 {
     command!()
-        .about("Convert CSV file to RDF using SPARQL")
+        .about("oxi-gen: Convert CSV file to RDF using SPARQL")
         .arg(
             Arg::new("delimiter")
                 .short('d')
@@ -547,7 +547,7 @@ where
         .get_matches_from(args)
 }
 
-pub fn configure_transform<I>(args: I) -> OxiTarql
+pub fn configure_transform<I>(args: I) -> OxiGen
 where
     I: IntoIterator<Item = String>,
 {
@@ -565,7 +565,7 @@ where
         }
     };
 
-    OxiTarql {
+    OxiGen {
         delimiter: matches.get_one::<String>("delimiter").unwrap().to_string(),
         tab: matches.get_flag("tab"),
         test: match matches.get_one::<u32>("test") {

src/main.rs

@@ -1,14 +1,14 @@
-use oxi_tarql::configure_transform;
+use oxi_gen::configure_transform;
 use std::{env, time::Instant};
 
 fn main() {
     // parse the supplied arguments
     let os_args: Vec<String> = env::args_os().map(|a| a.into_string().unwrap()).collect();
-    let mut tarql = configure_transform(os_args);
+    let mut transform = configure_transform(os_args);
 
     let start = Instant::now();
 
-    tarql.transform().expect("Transformation failed");
+    transform.transform().expect("Transformation failed");
 
     let duration = Instant::now().duration_since(start);
     eprintln!("Processing complete in {} seconds", duration.as_secs_f32());

tests/fixtures/escaped_chars.csv

@@ -0,0 +1,5 @@
+id,description
+1,"C:\\Users\\test\\path"
+2,"line1\\nline2\\ttab"
+3,"say \"hello world\""
+4,"mixed: C:\\path \"quoted\""

tests/fixtures/escaped_chars.rq

@@ -0,0 +1,9 @@
+prefix : <https://test.com/d/>
+construct {
+    ?row_iri
+        :description ?description ;
+        .
+}
+where {
+    BIND(tarql:expandPrefixedName(CONCAT(":row_", STR(?ROWNUM))) AS ?row_iri)
+}